AUTHORS
Aman Wadhwa, MD, MSPH, Chen Dai, MS2, Omar Ahmed, MBBCh3, Sandra Kessel4, Joshua Richman, MD, PhD5, Wei Shen, MD MPH MS6, Jennifer A Belsky, DO, MS7, Annalynn Williams, PHD8, Justine M. Kahn, MD, MS9, Sharon M Castellino Jr, MD, MSc10, Kara M. Kelly, MD11, Debra Friedman, MD MS12, Smita Bhatia, MD, MPH13
AFFILIATIONS
1 University of Alabama at Birmingham, Birmingham, Alabama, United States
2 University of Alabama at Birmingham, Birmingham, Alabama, United States
3 University of Alabama at Birmingham, Birmingham, Alabama, United States
4 Quality Assurance Review Center, Providence, Rhode Island, United States
5 University of Alabama at Birmingham, Birmingham, Alabama, United States
6 Columbia University, New York, New York, United States
7 Indiana University School of Medicine, Indianapolis, Indiana, United States
8 University of Rochester Medicine, Rochester, United States
9 Columbia University, New York, New York, United States
10 Emory University, Aflac Cancer and Blood Disorders Center, Children’s Healthcare of Atlanta, Atlanta, Atlanta, United States
11 Roswell Park Comprehensive Cancer Center, Buffalo, United States
12 Vanderbilt University School of Medicine and Vanderbilt-Ingram Cancer Center, Nashville, United States
13 University of Alabama at Birmingham, Birmingham, Alabama, United States
ABSTRACT
Introduction: Exposure to cytotoxic chemotherapy is associated with acute treatment-related toxicities (TRT: hematologic; non-hematologic) in children/adolescents with Hodgkin lymphoma (HL). These toxicities may lead to reduction or omission in therapy, potentially impacting disease-free survival, presenting a need to identify novel mechanisms underlying the observed inter-individual variability in TRT risk. Among adults with cancer, computed tomography (CT)-based body composition (skeletal muscle and adipose tissue mass) is associated with TRT risk. Whether body composition at diagnosis in children/adolescents with HL is associated with TRT remains unknown.
Methods: The current report included 1,165 patients with intermediate-risk HL enrolled on Children’s Oncology Group (COG) therapeutic trial (AHOD0031) for whom digitized pre-treatment abdominal CT scans were available. Patients received one of three treatment regimens: ABVE-PC (doxorubicin, bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide) alone, ABVE-PC plus IFRT (involved-field radiation therapy) or ABVE-PC, DECA (dexamethasone, etoposide, cisplatin and cytarabine) and IFRT; chemotherapy dosing was not capped based on body weight except for vincristine. Using a commercially available software (sliceOmatic), two consecutive abdominal CT slices at L3 level were identified and skeletal muscle area and total adipose tissue area was calculated and adjusted for height, yielding skeletal muscle index (SMI) and total adipose tissue index (TATI). SMI and TATI were divided into quintiles (Q1 [lowest] to Q5 [highest]) using sex-specific distribution. Body mass index (BMI) was calculated using patient height and weight at diagnosis and converted into age- and sex-specific percentiles to assign patients into the following BMI categories: underweight/normal weight (<85th percentile), overweight (85th-94th), obese (≥95th). COG Statistics and Data Center provided relevant demographic and clinical data, as well as CTCAE v4.0 graded TRTs over the entire treatment period. The primary outcome was cumulative count of grade ≥4 hematologic toxicities and grade ≥3 non-hematologic toxicities (excluding allergic, psychiatric, constitutional toxicities). The association between body composition and TRT, and between BMI and TRT was examined using negative binomial regression including log(time on study) as an offset to account for variation in the number of chemotherapy cycles. The following covariates were considered: age at diagnosis, sex, race/ethnicity, stage, bulk disease, ‘B’ symptoms, therapeutic exposures (ABVE-PC alone, ABVE-PC plus IFRT; ABVE-PC, DECA and IFRT) and were selected for inclusion in multivariable models using LASSO regularization to account for collinearity.
Results: Median age at diagnosis was 15.2y (range, 1.9-21.8); 53% were male and 64.5% were non-Hispanic White. Grade ≥4 hematologic toxicities were observed in 64.7% of patients and grade ≥3 non-hematologic toxicities in 46.4%.
Grade ≥4 hematologic TRT: Patients in SMI Q5 group (highest muscle mass) were significantly less likely to experience grade ≥4 hematologic toxicities (incidence rate ratio [IRR]=0.76, 95%CI=0.62-0.94, P=0.01; reference: SMI Q1). We did not observe associations between grade ≥4 hematologic toxicities and TATI quintiles or BMI categories.
Grade ≥3 non-hematologic TRT: Higher SMI quintiles were associated with lower rates of grade ≥3 non-hematologic toxicities (Q4: IRR=0.68, 95%CI=0.50-0.91, P=0.01; Q5: IRR=0.70, 95%CI=0.52-0.95, P=0.02; ref=Q1). In contrast, higher TATI quintiles (higher adipose tissue mass) were associated with higher rates of grade ≥3 non-hematologic toxicities (Q3: IRR=1.55, 95%CI=1.15-2.08, P=0.004; Q4: IRR=1.51, 95%CI=1.12-2.04, P=0.007; Q5: IRR=1.81, 95%CI=1.32-2.48, P<0.001; ref=Q1). Compared to underweight/normal weight BMI at diagnosis, overweight BMI was associated with higher rates of grade ≥3 non-hematologic toxicities (IRR=1.42, 95%CI=1.12-1.80, P=0.004) but obese BMI was not (IRR=1.14, 95%CI=0.90-1.45, P=0.27).
Conclusions: Assessment of imaging-based body composition can identify patients at greater risk of acute treatment-related toxicities. These findings could inform future TRT mitigations strategies.
DISCLOSURES
Aman Wadhwa, MD, MSPH: No relevant conflicts of interest to declare.